H.264運(yùn)動(dòng)估計(jì)算法在Imagine流處理器上的加速實(shí)現(xiàn):Imagine 是一款媒體處理的流體系結(jié)構(gòu)芯片，它提供三級(jí)帶寬層次，能夠充分開(kāi)發(fā)數(shù)據(jù)局域性和并行性，獲得很高的性能。H.264 是目前最新的數(shù)字視頻編碼標(biāo)準(zhǔn)，它性能優(yōu)越卻計(jì)算復(fù)雜。并且，視頻圖像具有自然的流特性，空間局域性很好而時(shí)間相關(guān)性也有限。因此，我們認(rèn)為H.264 編解碼過(guò)程能夠利用流處理來(lái)提高其性能。本文將H.264 編碼中耗時(shí)較大的運(yùn)動(dòng)估計(jì)算法映射在Imagine 模擬平臺(tái)上，實(shí)驗(yàn)結(jié)果表明，QCIF 圖像編碼效率可達(dá)438fps，遠(yuǎn)高于實(shí)時(shí)要求。與V1304 H.264 編碼器的25fps 相比，加速比達(dá)到17.5 倍，
流處理的加速效果是非常明顯的。由此證明，H.264 是適于Imagine 實(shí)現(xiàn)的。

Abstract: Imagine is a stream-based prototype chip designed for media processing. It uses a
three-level bandwidth hierarchy to exploit available parallelism and data locality. It has good
performance in media processing. H.264 is the newest digital video coding standard. It can
achieve high coding efficiency at the cost of complex computation. In addition, video pictures
have natural stream features, such as good special locality and limited temporal dependency. So it is thought that H.264 can increase its coding efficiency by stream processing. An accelerated implementation of motion estimation, which is the most time-consuming part in H.264 coding framework, on Imagine stream processor is presented in this paper. Experimental results show that the coding efficiency for QCIF format can be up to 438fps and meet real-time need. Compared with V1304 H.264 encoder, the speedup rate is 17.5x. The accelerated effect of stream processing is obvious. It proves that H.264 coding is suitable for implementation on Imagine.
關(guān)鍵詞：Imagine 流處理器；H.264 視頻編碼；運(yùn)動(dòng)估計(jì)
Key words: Imagine stream processor; H.264 video coding; motion estimation